INDEX

Explanations

technical terms and phrases related to science, medicine, and engineering

oai_token-act-pair · gpt-4o-mini Triggered by @bot

names of things

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Administrativna

-0.56

 ivelany

-0.44

 Italijani

-0.41

 الحره

-0.40

imakasih

-0.38

olla

-0.38

quema

-0.37

也好

-0.36

 Италијани

-0.36

 segura

-0.36

POSITIVE LOGITS

 surla

0.60

OGND

0.60

DockStyle

0.59

AddTagHelper

0.57

 autorytatywna

0.56

 CreateTagHelper

0.55

Попис

0.54

expandindo

0.52

tagHelperRunner

0.51

aarrggbb

0.51

Activations Density 0.071%

technical terms and phrases related to science, medicine, and engineering

names of things

No Comments

No Known Activations